CORROSION SCIENCE SECTION CORROSION—Vol. 70, No. 3 283 Submitted for publication: May 4, 2013. Revised and accepted: October 8, 2013. Preprint available online: October 17, 2013, doi: http://dx.doi.org/10.5006/1003. ‡ Corresponding author. E-mail: kamachi@igcar.gov.in. * Indira Gandhi Centre for Atomic Research, Kalpakkam – 603102, India. Electrochemical Noise Analysis of Pitting Corrosion of Type 304L Stainless Steel Girija Suresh* and U. Kamachi Mudali ‡, * ABSTRACT Electrochemical current and potential noise were simultane- ously acquired from Type 304L stainless steel (UNS S30403) in 0.05 M ferric chloride (FeCl 3 ) using a three-electrode con- iguration . Power spectral, statistical, and wavelet analyses have been used to know the uniqueness of the parameters proposed for the identiication of various types of corrosion processes. Roll-off slopes derived from power spectral analy- sis and statistical parameters such as standard deviation, localization index, and kurtosis corroborated with pitting as the corrosion mechanism. Energy distribution plots (EDP) ob- tained from wavelet analysis of current noise was found to be useful to derive mechanistic information on the progress of corrosion. Discrete wavelet transform was used to decompose the signals into a D 1 , D 2 , D 3 …D 8 , S 8 set of coeficients. The EDP showed that the contribution from the medium time scale crystal, D 5 , prevailed over the smaller time scale crystals and larger time scale crystals during the initial stages of immer- sion. With an increase in the time of immersion, the energy deposition on the larger time scale crystals increased and the maximum energy was concentrated on the D 8 crystals, indi- cating that the dominant process occurring on the specimen surface was stable pitting. The results of the investigation are detailed in the paper. KEY WORDS: electrochemical noise, pitting corrosion, power spectra, statistical analysis, wavelet transform INTRODUCTION Electrochemical noise has progressed a long way since its description by Iversion in 1960s. 1 Extensive research undergone in the last 30 years has elevated the technique to a relatively matured state. The time- dependant luctuation of current and potential dur- ing corrosion process have been used to indicate the type of attack and the rate. An important credential of this technique is to identify and quantify localized corrosion where other techniques are substantially less effective; yet, many of the proposed noise results are debatable and can lead to ambiguous results. For uniform corrosion, the methods of noise analysis using noise resistance and impedance are quite well established; however, the understanding is limited for localized corrosion. Electrochemical noise parameters have been deduced by various investigators to under- stand localized corrosion. 2 Some of the methods that are proposed in literature utilize power spectrum, 2-7 , statistical parameters such as skewness, kurtosis, 8-10 localization index, 2,11 bispectrum, or estimation of the intensity of characteristic transient occurrence in voltage, or current records 12 for monitoring pitting corrosion. Cottis, et al., 13-14 opines that the charac- teristic charge and frequency provide more informa- tion about localized corrosion and have associated large charge and low frequency for pitting corrosion. Wavelet transform is relatively a new mathematical tool that has gained popularity for analyzing elec- trochemical noise signals. 15-18 Two types of wavelet transform have been developed: the continuous wave- let transforms (CWT) and discrete wavelet transforms ISSN 0010-9312 (print), 1938-159X (online) 14/000049/$5.00+$0.50/0 © 2014, NACE International